Paul Stansifer

`match_dotdotdot` is now gone. I had to change the semantics of dotdotdots (now they are always tuples, rather than working inside any kind of repeated construct) to avoid needing to...

(In ccead10, I restored the old semantics of dotdotdot, but without using `pre_match`. On the other hand, I introduced something else to `ast_walk::walk` that is in obvious need of refactoring...)

I wonder if, instead of immediately doing the "push-through" in `pre_match`, we could just have the walk result in a variable (or underdeterminedness) equalling some other thing, and *outside* the...

(also, `walk_ctxt` is sometimes used for `LazyWalkReses`s, but I think that is probably confuseable with `context_elt`)

The commit to hopefully back out is ed67b9e.

Okay, there's no way that'll back out cleanly, but the addition of `Language` should make it pretty straightforward to remove.

Okay, the problem with `forall T. T` is that it doesn't have kind annotations (and we don't have a kind-checker yet (#3)). So I think we can get away with...

This is the same thing as what the `MuProtect` walk currently does.

Named repeats (actually, repeat groups; they don't need to be named) can't go away. For `Custom` walks, we can rely on the user naming all of the repeats that need...

The proposed method isn't a complete fix; what if we have `Star(Star(Named(...))`? You might have 5 copies at the first layer, each of which has zero copies, or you might...